Tables A1-A3 are only available in electronic form at http://www.edpsciences.org and at the CDS via anonymous ftp to cdsarc.u-strasbg.fr (130.79.128.5) or via http://cdsweb.u-strasbg.fr/cgi-bin/qcat?J/A+A/392/1129

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

The analysis technique with stellar ratios was employed. The reason for this is that ratios are less sensitive to uncertainties in certain regions in the CMD, which might not be reproduced properly due to various reasons such as the input physics used for the calculations of the stellar evolutionary tracks or the transformations from the theoretical to the observational plane.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...ZVAR

Note that an older and modified version of HRD-ZVAR is actually the program used to generate the Bertelli et al. (1994) isochrones. HRD-ZVAR is free of the interpolation difficulties as reported by Olsen (1999).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... exponential

The SFR is only fixed for the time being to the adopted exponential shape.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...UBVRI

Note that the IR photometric system is based on an "average'' photometric system as described by Bessell & Brett (1988). Proper transformations ought to be applied to the actual photometric system prior to any astrophysical interpretation of the results.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... accuracy

The 1 or 2 number of digits encoding accuracy maps the parameter via an integer value, either 0-9 or 0-99, to a range of floating point values by a controlling "normalization'' function, see Charbonneau (1995) for additional details.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... survive

This is irrespective if the local, fittest individuals are weaker than the weakest individual in the global population or if the locally remaining offspring are fitter than the fittest individual in the globalpopulation.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... matching

This option is feasible due to the increment of the present day computational speed.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... smaller

For our testcase, as described in Sect. 4, $F_\chi$ ranges from 0.7-1.0 for all observed data points with a matching synthetic point within a 3 $\sigma$ uncertainty ellipse. This corresponds to a goodness of fit parameter ranging from 0.49-1.0. It further indicates that it is justified to assume that the measurement errors are normally distributed.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... $F\!\la\!2\;$

This condition is comparable but not equivalent to the results obtained by Gallart et al. (1999). They demonstrated from a comparison with colour-magnitude bins that a good agreement between the input and recovered SFR(t) required a reduced chi-squared of $\chi_\nu^2\!\simeq\!2.0$ .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... $f_{\sigma,k}$

For both age and metallicity ranges the associated uncertainties denote for the lower and upper values the $-1\sigma$ and $+1\sigma$ boundary.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... chromosome

The genetic information is currently located on one chromosome. Individuals with two chromosomes might be considered as a future extension. A two-chromosome approach has the advantage that certain genetic information can remain present in a recessive form.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... individuals

To avoid confusion the term group is used instead of the biological term population, because each individual in the group is actually a synthetic stellar population.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... Linux

Red Hat $^{\bigcirc\!\!\!\!{\tt R}}$ is a registered trademark of Red Hat Software, Inc. and Linux $^{\bigcirc\!\!\!\!{\tt R}}$ is a registered trademark of Linus Torvalds.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.